A processing data managing system includes: a processing device 11 (such as a sputtering device for manufacturing a magnetic disc) for repeating the same process for each cycle; a sampling unit 30 for collecting raw data on a processing condition in the processing device (such as a discharge condition); a calculation unit 100 for receiving the raw data, calculating the raw data according to a predetermined rule, and processing it as summary data expressing a characteristic point for each cycle (characteristic value: for example, average value, maximum value, minimum value, standard deviation, discharge time, and the like); a data storage unit 40 for storing the processed summary data in storage means; and a display/output unit 50 for chart-displaying the summary data stored in the storage means.
|
9. A processing device data managing method comprising:
a step of collecting raw data on process treatment conditions in a magnetic disk manufacturing apparatus that performs a same process treatment for each of repeated cycles,
a step of processing said raw data into a characteristic value for each of repeated cycles, and
a step of chart-displaying the characteristic value for each of repeated cycles,
the processing step obtaining the characteristic value for each of repeated cycles by converting the raw data for one of repeated cycles into the characteristic value representing characteristics of the raw data for one cycle,
the chart-displaying step chart-displaying the characteristic value for each of repeated cycles so as to visualize time-dependent changes of the characteristic values for the repeated cycles,
the magnetic disk manufacturing apparatus being configured to form magnetic disk thin films on the surface of a substrate by generation of discharge in a processing chamber maintained in a predetermined gas pressure atmosphere;
the raw data being such that the number of data per hour is 14400000 or more where at least the raw data on power supply output for generation of the discharge are collected at regular sampling intervals, one cycle is set to 6 seconds, and sampling is performed 40 times in each second;
the characteristic value being repeatedly calculated for each cycle, based on the raw data on the power supply output, on five items including a discharge ON time and an average value, a maximum value, a minimum value, and a standard deviation of the power supply output so that one cycle is represented by plot data at one point.
8. A processing device data managing method comprising:
a step of collecting raw data on process treatment conditions of film forming treatments for magnetic disks in a magnetic disk manufacturing apparatus that performs a same process treatment for each of repeated cycles,
a step of processing said raw data into a characteristic value for each of repeated cycles, and
a step of storing the characteristic value for each of repeated cycles in a storage,
the processing step obtaining the characteristic value for each of repeated cycles by converting the raw data for one cycle into the characteristic value representing characteristics of the raw data for one cycle,
the storing step storing the characteristic values for the repeated cycles in the storage so as to visualize time-dependent changes of the characteristic values for the repeated cycles,
the magnetic disk manufacturing apparatus being configured to form magnetic disk thin films on the surface of a substrate by generation of discharge in a processing chamber maintained in a predetermined gas pressure atmosphere;
the raw data being such that the number of data per hour is 14400000 or more where at least the raw data on power supply output for generation of the discharge are collected at regular sampling intervals, one cycle is set to 6 seconds, and sampling is performed 40 times in each second:
the characteristic value being repeatedly calculated for each cycle, based on the raw data on the power supply output, on five items including a discharge ON time and an average value, a maximum value, a minimum value, and a standard deviation of the power supply output so that one cycle is represented by plot data at one point.
2. A processing data managing system comprising:
a data processing section for receiving raw data on processing conditions of film forming treatments for magnetic disks collected in a magnetic disk manufacturing apparatus that performs a same process treatment for each of repeated cycles and processing aid raw data into a characteristic value for each of the repeated cycles, and
a display unit for chart-displaying the characteristic value for each of the repeated cycles,
said data processing section obtaining the characteristic value for each cycle by converting the raw data for one cycle into the characteristic value representing characteristics of the raw data for one cycle,
said display unit chart-displaying the characteristic values for the repeated cycles so as to visualize time-dependent changes of the characteristic values for the repeated cycles,
the magnetic disk manufacturing apparatus being configured to form magnetic disk thin films on the surface of a substrate by generation of discharge in a processing chamber maintained in a predetermined gas pressure atmosphere;
the raw data being such that the number of data per hour is 14400000 or more where at least the raw data on power supply output for generation of the discharge are collected at regular sampling intervals, one cycle is set to 6 seconds, and sampling is performed 40 times in each second:
the characteristic value being repeatedly calculated for each cycle, based on the raw data on the power supply output, on five items including a discharge ON time and an average value, a maximum value, a minimum value, and a standard deviation of the power supply output so that one cycle is represented by plot data at one point.
1. A processing data managing system comprising:
a data processing section for receiving raw data on processing conditions of film forming treatments for magnetic disks collected in a magnetic disk manufacturing apparatus that performs a same process treatment for each of repeated cycles and processing said raw data into a characteristic value for each of the repeated cycles, and
a data storage unit for storing the characteristic value for each of the repeated cycles in a storage,
said data processing section obtaining the characteristic value for each of the repeated cycles by converting the raw data for one cycle into the characteristic value representing characteristics of the raw data for one cycle,
said data storage unit storing the characteristic values for the repeated cycles in the storage so as to visualize time-dependent changes of the characteristic values for the repeated cycles,
the magnetic disk manufacturing apparatus being configured to form magnetic disk thin films on the surface of a substrate by generation of discharge in a processing chamber maintained in a predetermined gas pressure atmosphere;
the raw data being such that the number of data per hour is 14400000 or more where at least the raw data on power supply output for generation of the discharge are collected at regular sampling intervals, one cycle is set to 6 seconds, and sampling is performed 40 times in each second;
the characteristic value being repeatedly calculated for each cycle, based on the raw data on the power supply output, on five items including a discharge ON time and an average value, a maximum value, a minimum value, and a standard deviation of the power supply output so that one cycle is represented by plot data at one point.
3. A processing data managing system comprising:
a data processing section for receiving raw data on a processing conditions of film forming treatments for magnetic disks collected in a magnetic disk manufacturing apparatus that performs a same process treatment for each of repeated cycles and processing said raw data into a characteristic value for each of the repeated cycles,
a data storage unit for storing the characteristic value for each of the repeated cycles in a storage, and
a display unit for chart-displaying the characteristic value for each of the repeated cycles output from said data processing section or the characteristic value for each of the repeated cycles stored in said storage,
said data processing section obtaining the characteristic value for each of the repeated cycles by converting the raw data for one cycle into the characteristic value representing characteristics of the raw data for one cycle,
said data storage unit storing the characteristic values for the repeated cycles in the storage so as to visualize time-dependent changes of the characteristic values for the repeated cycles,
said display unit chart-displaying the characteristic values for the repeated cycles so as to visualize time-dependent changes of the characteristic values for the repeated cycles,
the magnetic disk manufacturing apparatus being configured to form magnetic disk thin films on the surface of a substrate by generation of discharge in a processing chamber maintained in a predetermined gas pressure atmosphere;
the raw data being such that the number of data per hour is 14400000 or more where at least the raw data on power supply output for generation of the discharge are collected at regular sampling intervals, one cycle is set to 6 seconds, and sampling is performed 40 times in each second;
the characteristic value being repeatedly calculated for each cycle, based on the raw data on the power supply output, on five items including a discharge ON time and an average value, a maximum value, a minimum value, and a standard deviation of the power supply output so that one cycle is represented by plot data at one point.
4. A processing data managing system according to any of
said characteristic value for each of the repeated cycles is data reduced in size with respect to the raw data for each of the repeated cycles.
5. A processing system comprising:
a magnetic disk manufacturing apparatus that performs a same process treatment for each of repeated cycles,
a raw data collecting unit for collecting raw data on processing conditions in said processing device, and
a processing data managing system according to any of
6. A processing system according to
said magnetic disk manufacturing apparatus is provided with a plurality of electrodes in the processing chamber for discharge generation, and
said raw data collecting unit collects raw data on a power supply output for discharge generation for each of said electrodes.
7. A processing system according to
said magnetic disk manufacturing apparatus comprises a plurality of processing chambers for forming thin films on the substrate in a predetermined order, and
said raw data collecting unit collects raw data on a power supply output for discharge generation for each of said processing chambers.
|
|||||||||||||||||||||||||||||||||||||||
This invention relates to a processing data managing system that is mainly applied to a manufacturing site or the like for high recording density magnetic disks, a processing system including such a processing data managing system, and a data managing method for a processing device or apparatus such as a magnetic disk manufacturing apparatus.
In recent years, the increase in magnetic disk recording density has been remarkable. A high recording density magnetic disk is manufactured through processes of forming a plurality of magnetic layers over the surface of a substrate. Since an underlayer and a protective layer are respectively provided under and over each of the magnetic layers, the magnetic disk is manufactured through a number of film forming processes. Therefore, in the manufacture of the magnetic disk, use is made of a processing device having a structure with a number of processing chambers arranged for carrying out the respective processes.
For example, when manufacturing one magnetic disk, a disk (substrate) is caused to pass through the number of processing chambers in order so as to be subjected to a predetermined process treatment in each of the processing chambers. In each processing chamber, one-time process treatment is given as one cycle and the same process treatment is repeated each time a disk passes therethrough, and, after the treatment, the disk is sent out to the next processing chamber. Then, a magnetic disk having gone through all the processes is sent out from the final processing chamber.
A sputtering device or apparatus is mainly used as the manufacturing apparatus (processing device) for such high recording density magnetic disks. The sputtering apparatus generates glow discharge by introducing a discharge gas into a vacuum and applying the power to electrodes so as to form a thin film of a target metal on the surface of a disk by collision of ions in a plasma generated by the glow discharge.
In each of processing chambers of the sputtering apparatus, there is a case where film formation is performed on one surface of a single disk, a case where film formation is performed on both surfaces of a single disk, a case where film formation is performed on one surface of each of two disks, a case where film formation is performed on both surfaces of each of two disks, or the like. For example, when simultaneously performing film formation on both surfaces of each of two disks in one processing chamber, it is necessary to provide four cathode electrodes for generating discharge and to apply the power to each of the electrodes.
In the meantime, when manufacturing magnetic disks of this type, it is important to store histories of the manufacturing processes. This is because if a defective product is produced due to occurrence of some abnormality in the manufacturing process, it is possible to pursue a cause thereof later and to promptly correct it.
At a conventional magnetic disk manufacturing site, process-treatment condition data (discharge condition, gas pressure, substrate temperature, etc.) are sampled and the sampled raw data are, as they are, displayed in a graph or the like or stored in storage means. Then, if a defective product is produced, the raw data extracted according to an index of a disk or a lot are, for example, displayed in a graph on a display screen in time sequence, thereby enabling an analysis while observing waveforms in the graph.
This manufacturing system comprises a processing device (sputtering apparatus) 10 for sputtering a disk being a workpiece W, a control unit 20 for controlling the processing device 10, a sampling unit (raw data collecting means) 30 for collecting raw data on processing conditions (power supply output and gas pressure for generating discharge, substrate temperature, etc.) in the sputtering, a data storage unit (data storage means) 40 for storing the collected data in recording means, and a display/output unit (display means) 50 for displaying in a chart or outputting to another output means the collected data or the stored data according to need.
The processing device 10 is provided with a processing chamber 11 for carrying out a necessary process treatment, a processing unit 12 for carrying out the treatment, a sensor 13 for monitoring the processing state, and so on. Raw data output from the processing unit 12 and the sensor 13 are input into the sampling unit 30. The data output from the sensor 13 and the processing unit 12 may be analog data or digital data. In the sampling unit 30, the raw data are properly converted into a form (digital data form) readable by the data storage unit 40 and the display/output unit 50. Thereafter, the converted data are stored or displayed/output.
The actual magnetic disk manufacturing processing is carried out through a plurality of processes in sequence. Therefore, in the processing device 10, a plurality of processing chambers 11 are arranged so that thin films can be formed on the surface of a disk in a predetermined order. In each of the processing chambers 11, the same process treatment is repeated for each cycle. Since, in this case, the processing device 10 is configured to form magnetic disk thin films on the surface of a disk (substrate) by generation of discharge in the processing chambers 11 each maintained in a predetermined gas pressure atmosphere, respective treatments for forming the thin film are carried out in order in each processing chamber 11.
The sampling unit 20 collects raw data on processing conditions in the processing device 10 and, herein, collects raw data on processing conditions in thin film forming treatments for a number of magnetic disks. For example, it collects raw data such as power supply outputs (power, voltage, current) for generating discharge, gas pressures in the processing chambers 11, and disk temperatures at regular sampling intervals.
Incidentally, in such treatments, acquisition (sampling) of raw data is often performed a plurality of times in the treatment of one process (cycle) in each processing chamber 11. This is for examining time-dependent changes of various parameters in one cycle in detail and, particularly, with respect to such a parameter that largely affects the processing results, it is necessary to carefully observe a change in one process and thus the sampling is performed this way.
When the data thus collected are displayed in time sequence, there is obtained, for example, a graph like
In the case of this example, the respective components correspond to the following specific means.
(Name)
(Example of Specific Means)
processing device 10
sputtering apparatus
processing region
inside of chamber
(processing chamber 11)
processing unit 12
sputtering cathode,
discharge power supply, etc.
sensor 13
pressure gauge etc.
control unit 20
PLC
sampling unit 30
A/D converter
display/output unit 50
computer
data storage unit 40
hard disk drive
In the manufacture of high density recording magnetic disks using the manufacturing system described above, the requirement for the film forming quality (film thickness and film property) is quite strict and, therefore, it is necessary to shorten the sampling interval in one cycle or increase the number of sampling items (parameters).
Specifically, for example, one cycle is set to 6 seconds and sampling is performed 40 times in one second. As detection items (parameters), there are power, voltage, and current with respect to the power supply output and, in addition, there are gas pressure and disk temperature. When simultaneously processing two front and back surfaces of each of two disks in one processing chamber 11, the number of cathodes is four. Therefore, there are 3 items×4=12 items with respect to the power supply output and, adding one item of gas pressure and two items of temperature of the two disks, it is necessary to obtain raw data on the 15 items in total in each processing chamber 11. Thus, the amount of raw data collected over the entire processing device 10 having the number of processing chambers 11 becomes quite enormous.
Besides, since waveforms of raw data are merely graph-displayed as they are in the conventional system, there has been a case where it is difficult to perform an analysis and thus to find a problem.
Under these circumstances, it is possible to point out that the conventional magnetic disk manufacturing system has the following problems (1), (2), and (3).
(1) There is a case where it is difficult to detect abnormality even by observing waveforms of raw data on a screen and, particularly because of enormous data, there is a case where if a reduced display is performed over a long-term span, it is possible to see the whole, but is not possible to read details, while, if an enlarged display is performed in a short-term span, it is possible to see details, but is difficult to read a long-term change or the like.
That is, raw data are plotted in a graph as they are in the conventional manufacturing system.
Accordingly, by the current data managing method, when abnormality occurs in data, it is difficult to find an abnormality occurring portion. That is, it is difficult to compare one-cycle waveforms and it is difficult to confirm a waveform long-term change.
(2) An acceptance/rejection judgment is difficult only by observing raw data waveforms.
That is, when using this type of raw data, a method is often adopted in which a human being observes waveforms to judge acceptance or rejection of a treatment. In this case, if the cycle is about one cycle per day to about one cycle per hour, it is easy for a human being to confirm each time, but if faster, i.e. when one cycle is as short as 6 seconds as in this example, it is necessary to confirm a raw data waveform per 6 seconds by visual observation and thus it is difficult for a human being to continue the observation for a long time.
As a measure for this, it is considered to provide upper and lower limit alarm values and to generate an alarm when one of them is exceeded. When a plurality of processing devices are used in parallel, it is practically impossible to judge acceptance or rejection by visual observation and, therefore, such an alarm function is essential. However, with this alarm function, it is not possible to grasp a change within a range not exceeding the alarm set values.
(3) Since the amount of raw data becomes enormous, data storage is difficult.
That is, in the foregoing conventional manufacturing system, if raw data are successively stored into the data storage unit 40, the data amount becomes enormous. Therefore, it is necessary to provide a large-capacity data storage unit 40 or it is necessary to record data into another backup medium per fixed period of time and then to erase the stored contents to ensure a free area. Although such raw data include useful information, it is not that the raw data itself at each point includes useful information. Useful information is a certain kind of information included in the entire waveform and is not individual raw data themselves. That is, it is wasteful to store the raw data as they are.
Therefore, in consideration of the above circumstances, it is an object of this invention to provide a processing data managing system that can solve the above problems by processing raw data into summary data (hereinafter may also be referred to as a characteristic value) representing a characteristic point for each cycle and then storing it, a processing system including such a processing data managing system, and a data managing method for a processing device such as a magnetic disk manufacturing apparatus.
According to structure 1 of this invention, there is provided a processing data managing system comprising:
data processing means for receiving raw data on a processing condition collected in a processing device that repeats the same process treatment for each cycle and calculating said raw data according to a predetermined rule to thereby process said raw data into summary data representing a characteristic point for each cycle, and
data storage means for storing the processed summary data in storage means.
According to structure 2 of this invention, there is provided a processing data managing system comprising:
data processing means for receiving raw data on a processing condition collected in a processing device that repeats the same process treatment for each cycle and calculating said raw data according to a predetermined rule to thereby process said raw data into summary data representing a characteristic point for each cycle, and
display means for chart-displaying the processed summary data.
According to structure 3 of this invention, there is provided a processing data managing system comprising:
data processing means for receiving raw data on a processing condition collected in a processing device that repeats the same process treatment for each cycle and calculating said raw data according to a predetermined rule to thereby process said raw data into summary data representing a characteristic point for each cycle,
data storage means for storing the processed summary data in storage means, and
display means for chart-displaying the summary data output from said data processing means or the summary data stored in said storage means.
According to structure 4 of this invention, there is provided a processing data managing system according to any of structures 1 to 3, wherein:
said raw data is raw data on processing conditions of thin film forming treatments for a number of workpieces, and
said summary data is data reduced in size by converting said raw data on the processing conditions of the thin film forming treatments according to a predetermined analysis point of view.
According to structure 5 of this invention, there is provided a processing data managing system according to structure 4, wherein:
said raw data is raw data on processing conditions of thin film forming treatments for a number of magnetic disks, and
said summary data is data reduced in size by converting said raw data on the processing conditions of the thin film forming treatments for the number of magnetic disks according to a predetermined analysis point of view.
According to structure 6 of this invention, there is provided a processing system comprising:
a processing device that repeats the same process treatment for each cycle,
raw data collecting means for collecting raw data on a processing condition in said processing device, and
the processing data managing system according to any of structure 1 to 5 for managing the raw data collected by said raw data collecting means.
According to structure 7 of this invention, there is provided a processing system according to structure 6, wherein:
said processing device is a thin film forming apparatus for forming a thin film on a surface of a substrate by generation of discharge in a processing chamber maintained in a predetermined gas pressure atmosphere,
said raw data collecting means collects, at regular sampling intervals, at least raw data on a power supply output for generating the discharge, and
said processing data managing system, based on said raw data on the power supply output, repeatedly calculates summary data, for each cycle, on at least one of items such as a discharge ON time and an average value, a maximum value, a minimum value, and a standard deviation of the power supply output and displays or stores said summary data.
According to structure 8 of this invention, there is provided a processing system according to structure 7, wherein:
said processing device is provided with a plurality of electrodes in the processing chamber for discharge generation, and
said raw data collecting means collects raw data on a power supply output for discharge generation for each of said electrodes.
According to structure 9 of this invention, there is provided a processing system according to structures 7 or 8, wherein:
said processing device comprises a plurality of processing chambers for forming thin films on the substrate in a predetermined order, and
said raw data collecting means collects raw data on a power supply output for discharge generation for each of said processing chambers.
According to structure 10 of this invention, there is provided a processing system according to any of structures 7 to 9, wherein:
said thin film forming apparatus is a magnetic disk manufacturing apparatus.
According to structure 11 of this invention, there is provided a processing device data managing method comprising:
a step of collecting raw data on a process treatment condition in a processing device that repeats the same process treatment for each cycle,
a step of calculating the collected raw data according to a predetermined rule to thereby process said raw data into summary data representing a characteristic point for each cycle, and
a step of storing the processed summary data in storage means.
According to structure 12 of this invention, there is provided a processing device data managing method comprising:
a step of collecting raw data on a process treatment condition in a processing device that repeats the same process treatment for each cycle,
a step of calculating the collected raw data according to a predetermined rule to thereby process said raw data into summary data representing a characteristic point for each cycle, and
a step of chart-displaying the processed summary data.
According to structure 13 of this invention, there is provided a processing device data managing method according to structure 11 or 12, wherein:
said process treatment in said processing device is a thin film forming treatment for forming a magnetic disk thin film on a surface of a substrate by generation of discharge in a processing chamber maintained in a predetermined gas pressure atmosphere.
(1) By using summary data, it is possible to easily grasp a change in processing state for each cycle. That is, it is possible to easily provide an analysis edge in a form suitable for analysis.
(2) By using summary data, it is possible to perform an acceptance/rejection judgment that is difficult with raw data, which thus is useful for preventing the occurrence of a defective product or judging a failure.
(3) By storing summary data, it is possible to reduce the data storage amount. Therefore, it is highly effective for the manufacture of high recording density magnetic disks where an extremely large amount of raw data are produced.
Hereinbelow, an embodiment of this invention will be described.
At first, to give a schematic description, the raw data described above include important information indicative of the stability of discharge. A point of this invention is how well to extract and display such data to thereby reduce the total data size.
For example, from the power supply output raw data, it is possible to calculate a discharge time, the average power during a discharge period of time, and so on. That is, it is considered that the discharge characteristics can be represented by a small number of data extracted from one-cycle raw data.
Naturally, the raw data include detailed information not included in these extracted data, for example, discharge leading and trailing states, the power stability during discharge, and so on. However, if some numerical values are obtained by processing the raw data as described above, these values can be plotted instead of the raw data and, by extensively continuing the plotting, it is possible to visually grasp time-dependent changes in discharge time and average power, for example. In this case, since one cycle is represented by plot data at one point, it is possible to display a relatively long-time tendency on one screen and, further, it is also possible to largely reduce the data size.
The discharge leading state and so on each can also be represented by one numerical value through proper calculation, but herein, some values (in this specification, referred to as “summary data” or “characteristic values”) representing the discharge characteristics other than those are extracted from the raw data, thereby precisely visualizing time-dependent changes of the data and reducing the data size.
In this embodiment, as characteristic values representing such discharge characteristics, items of
are employed. In addition thereto, it is considered to also employ the number of times of abnormal discharge (counting rapid output drops).
This manufacturing system comprises a processing device (sputtering apparatus) 10 for sputtering a disk being a workpiece W, a control unit 20 for controlling the processing device 10, a sampling unit (raw data collecting means) 30 for collecting raw data on processing conditions (power supply output and gas pressure for generating discharge, substrate temperature, etc.) in the sputtering, a calculation section 100 as data processing means for receiving the collected raw data and calculating the raw data according to a predetermined rule to thereby process the raw data into summary data representing a characteristic point for each cycle, a data storage unit (data storage means) 40 for storing the processed summary data in recording means, and a display/output unit (display means) 50 for displaying in a chart or outputting to another output means the collected data or the stored data according to need.
In terms of a data managing method, the sampling unit 30 executes a step of collecting raw data on the processing conditions, the calculation section 100 executes a step of calculating the raw data according to the predetermined rule to thereby process the raw data into summary data representing a characteristic point for each cycle, the data storage unit 40 executes a step of storing the summary data in the storage means, and the display/output unit 50 executes a step of chart-displaying the summary data stored in the storage means.
Herein, a group comprising the calculation section 100, the data storage unit 40, and the display/output unit 50 can be separated from the processing device 10 side so as to be independent as a processing data managing system. Further, among them, it is possible to separate the calculation section 100 and the data storage unit 40 from the display/output unit 50. Separation means independence both spatially and temporally.
The summary data are data reduced in size by converting the raw data on the processing conditions of the sputtering (thin film forming treatment) according to the predetermined analysis point of view and are specifically, as described above, the discharge ON time and the power supply output average value, maximum value, minimum value, and standard deviation calculated on the basis of the power supply output.
The processing device 10 is provided with a processing chamber 11 for carrying out a necessary process treatment, a processing unit 12 for carrying out the treatment, a sensor 13 for monitoring the processing state, and so on. Raw data output from the processing unit 12 and the sensor 13 are input into the sampling unit 30. Data output from the sampling unit 30 are processed in the calculation section 100 and then input into the data storage unit 40 and the display/output unit 50. Naturally, it may be configured to display/output the raw data, in the form not processed in the calculation section 100, in/to the display/output unit 50.
As is clear from the block diagram of
The calculation section 100 converts raw data for one cycle into a value representing its characteristics (this is called a “characteristic value” and corresponds to “summary data” in claims). It is possible to define a plurality of characteristic values with respect to one-cycle raw data for one measurement item (parameter) and, herein, as an example, five characteristic values of average value, maximum value, minimum value, standard deviation, and discharge ON time are defined with respect to the discharge voltage.
The calculation section 100 starts calculation by an external or internal trigger signal, finishes the calculation by a calculation end trigger, and outputs obtained characteristic values. An example of internal triggers is as follows. That is, when input data satisfies a predetermined condition, the calculation is started and, when input data does not satisfy the condition, the calculation is finished. Specifically, as illustrating in
When values during processing (during discharge) and during non-processing (during non-discharge) differ from each other like the discharge voltage, data calculation can be automatically started and finished by internal triggers based on a threshold value. However, for a parameter whose value does not change between during processing and during non-processing or whose variation is greater than a difference between during processing and during non-processing, internal triggers cannot be used. In such a case, calculation is started in response to a threshold value of another parameter or calculation is started by a cycle-synchronous calculation start command from the control unit 20. As an example of such a parameter, there is a “gas pressure”. Although depending on the process conditions, the pressure takes substantially the same value both during processing and during non-processing and thus it is difficult to perform calculation based on a threshold value.
The characteristic values calculated in the calculation section 100 are output per cycle and the output characteristic values are properly sent to the data storage unit 40 and the display/output unit 50. A program (not illustrated) using a database can be used for data storage and display.
As an Example, the following conditions are assumed herein.
one cycle
6 seconds
processing content
thin film formation by sputtering
processing time
one second
processing unit
sputtering power supply
(output: voltage, current, power)
sensor
pressure gauge
sampling interval
40 points/sec
acquisition data
discharge power during sputtering
characteristic values
ON time, average value, maximum
value, minimum value, standard
deviation
threshold value
5 W
Data were processed under the above conditions to obtain characteristic values.
In the foregoing raw data graph of
On the other hand,
From
From these, it can be confirmed that the discharge time and the discharge power are both substantially constant over 250 cycles with respect to the power. In this manner, by plotting the characteristic values instead of the raw data, it is possible to easily confirm the stability of the repeated cycles.
However, such a difference tends to be overlooked by visual confirmation and, in the case of short-time cycles, the visual confirmation itself becomes difficult.
By paying attention to the standard deviation, it is considered that since the standard deviation takes a value equivalent to that (0.4 W) in the normal state at 26 to 50 disks, 76 to 100 disks, . . . , the cycles of 26 to 50 disks, 76 to 100 disks, . . . are normal. On the other hand, it is seen that the standard deviation is 0.6 to 0.8 in the cycles of 1 to 25 disks, . . . and thus the variation in discharge power is large in these cycles.
Therefore, it is possible to judge that the possibility is high that the processing device has a problem during this data acquisition. As a result of actual examination, it was found that another processing unit connected to this device performed a treatment per cycle of 25 disks and noise generated from a cable of a high-frequency power supply used therein affected the sputtering power supply. In this manner, by the conversion into the characteristic values, it is possible to clearly display the abnormality that is difficult to confirm by visual observation.
Next, characteristic value calculation examples will be given.
Various values can be considered as characteristic values. Herein, calculation methods will be described using relatively easy examples. Other than these, as described before, various statistical values of waveform rise time and discharge period of time, and so on can be considered. Further, other than the calculation methods given herein, there are considered to be a number of methods that can obtain the same results and, therefore, the methods are not particularly limited thereto.
<Average Value Avg>
When raw data becomes a threshold value or more, the raw data are successively placed into array variables C(n) and, when the raw data becomes less than the threshold value, the placement is finished. An average value (avg) is calculated from the placed m values.
Average Value=(C(1)+ . . . +C(m))/m
<Maximum Value Max>
When raw data becomes a threshold value or more, the raw data are placed into array variables D(n) and, when the raw data becomes less than the threshold value, the placement is stopped. Then, the magnitudes of the array variables are compared and the greatest value is set as a maximum value (max). For acquiring proper data, regions where raw data is not collected need to be provided in a constant period of time before and after the start/end of discharge.
<Minimum Value Min>
When raw data becomes a threshold value or more, the raw data are placed into array variables D(n) and, when the raw data becomes less than the threshold value, the placement is stopped. Then, the magnitudes of the array variables are compared and the smallest value is set as a minimum value (min). For acquiring proper data, regions where raw data is not collected need to be provided in a constant period of time before and after the start/end of discharge.
<Standard Deviation Dev>
When raw data becomes a threshold value or more, the raw data are successively substituted into array variables F(n) and, when the raw data becomes less than the threshold value, the substitution is finished. A standard deviation (dev) is calculated from the substituted m values. (formula omitted)
<ON Time Difference>
An ON time difference between two measurement items is calculated. A time instant when the item 1 exceeds a threshold value is put into a variable G and a time instant when the measurement item 2 exceeds a threshold value is put into H. The ON time difference is calculated by H−G. This is used for measuring a time from the power supply output ON to the start of actual discharge.
Depending on a characteristic value, there is a case where if values immediately after the discharge or immediately before the end thereof are collected, an appropriate value is not output (example: maximum value, maximum value). Therefore, for preventing the collection thereof, regions where data is not collected (regions to be ignored) may be preferably set as illustrated in
<Discharge ON Time (+Number of Times of Arc Generation)>
A time instant when raw data becomes a threshold value or more is recorded into a variable A as a discharge start time instant and a time instant when the raw data becomes less than the threshold value is recorded into a variable B as a discharge end time instant. A discharge time is calculated by the following calculation formula. Discharge Time=B−A.
In the above method, when an arc is generated so that the discharge voltage becomes zero, the measurement is finished at that time instant. Thereafter, when the arc disappears and the discharge starts again, data start to be collected again. Therefore, as it is, a time from final arc generation to the discharge end is derived.
For preventing this, it is effective to perform an AND operation with a power supply output ON command and to set A and B into an array.
For example, as illustrated in
When the output is started again after restoration from the arc, an arc restoration time instant is substituted into A(2) and, thereafter, an arc time instant and a restoration time instant are substituted into A(k) and B(k) per occurrence of an arc and k is incremented each time. Then, when the power supply ON signal is stopped and the output becomes less than the threshold value, the above operation is finished.
By performing the above processing, it is possible to calculate a discharge time in consideration of the arc and, further, it is possible to derive the number of times of the arc. However, in terms of judging the discharge with arc generation to be abnormal, it may be better not to incorporate such processing.
Actual Discharge Time=B(0)−A(0)−B(1)−A(1)+ . . . +B(k)−A(k)
Number of Times of Arc Generation=k
In each of the methods described above, use is made, as a trigger for calculation of one characteristic value (e.g. power), of a threshold value of the original raw data (power raw data). However, another raw data (e.g. voltage) may be used as a trigger. For example, the voltage is used as a calculation trigger for the power, and so on.
As described above, by storing, instead of raw data, characteristic values (summary data) in the data storage unit 40, the data storage amount can be reduced and therefore it is advantageous when collecting data over a long period of time or measuring a large number of items.
In the above Example, sampling is performed at 40 points per second and, therefore, when storing raw data, the data amount becomes 144000 data per hour. In the magnetic disk manufacturing sputtering process or the like, a plurality of processing devices are connected together and, further, there are several tens of measurement items in each processing chamber, and therefore, the number of total measurement items often exceeds 100.
For example, when data on 100 items are sampled at 40 points per second, the data amount becomes 691 MB per day, 20 GB per month, and 252 GB per year in the case of the storage of raw data. In contrast, in the case of the storage of characteristic values, even the data for one year are reduced to 52 MB and thus the handling of data becomes extremely easy as compared with the raw data.
The data amounts are compared. The conditions are as follows.
number of times of sampling
40
points/sec
cycle time
6
seconds
measurement item
100
items
data byte
2
byte/data
number of characteristic values
5
items
<Comparison Table>
(In Case of
(In Case of
Characteristic
Raw Data)
Values)
number of data per cycle
240/cycle
5/cycle
number of cycles per hour
600
cycles
600
cycles
per point per hour
144000
600
number of data per hour
14400000
3000
number of bytes per hour
28.8
MB
6
KB
number of bytes per day
691.2
MB
144
KB
number of bytes per 30 days
20.7
GB
4.3
MB
number of bytes per year
252.3
GB
52.6
MB
Next, a description will be given of a way of using (extendability of) the foregoing embodiment.
As described above, by converting raw data into characteristic values, it is possible to easily perform abnormality detection that conventionally relied on visual observation. For example, it is very difficult to confirm a discharge time using raw data, but it can be easily confirmed using a characteristic value called a discharge ON time.
In addition to the foregoing configuration, there may be preferably provided reference value setting means for setting an upper or lower limit reference value with respect to each of characteristic values (summary data), abnormality judging means for judging abnormality based on comparison between the reference values and the characteristic values, and alarm generating means for generating an alarm according to a result of the judgment by the abnormality judging means. For example, if upper/lower limit alarms are set with respect these characteristic values, abnormality can be effectively discovered. Also with respect to variation during discharge, if upper/lower limit alarms are set for the characteristic values such as the standard deviation, the maximum value, and the minimum value, it is possible to use them for an acceptance/rejection judgment or the like. Even with respect to a somewhat complicated waveform, if the characteristics of the waveform can be well converted into numerical values, it is possible to set an upper/lower limit alarm and thus to perform abnormality detection that cannot be realized with an upper/lower limit alarm of raw data. In this manner, if the characteristic value is used, the upper/lower alarm can be effectively utilized, which is beneficial. In this case, structurally, there may be preferably provided means for judging whether or not a characteristic value (summary data) output from the calculation section 100 falls within a proper range and means for generating an alarm when the characteristic value is judged to deviate from the proper range. By adding all or part of the above configuration to the basic configuration, it is possible to effectively notify and discover abnormality.
Further, as another example, the data management using the characteristic values can be associated with the raw data.
Next, a description will be given of a case where it is extended as a production management system.
If the foregoing alarm information is input and stored in a database along with the characteristic values, it is possible to easily examine the behavior of the characteristic values at the time of the generation of the alarm. As a production management system in that case, it is preferable to have the following functions.
(a) When the date and a disk number are input, a table of characteristic values of a disk concerned is displayed.
(b) When there is an alarm, the alarm contents are also displayed.
(c) When the date and a cassette number are input, a table of characteristic values of all disks of a cassette concerned is displayed. Further, the alarm contents are also displayed.
(d) When a period of time and a chamber number are input, a trend graph of characteristic values of a chamber concerned is displayed. Further, an alarm generated place is marked and the alarm contents are displayed.
(e) When a period of time and a chamber number are input, characteristic values of a chamber concerned are output in the form of an Excel file or a CVS file. Further, alarm generation data is also output.
(f) A recipe change history and actual values are displayed in a graph and a portion with a deviation of a certain value or more is highlighted.
These functions can all be realized by jointly using a database.
When the production management system having the above functions is constituted, it is possible to achieve an improvement in quality management and a quick response by free information exchange between a product user and a product manufacturer through communication means such as the Internet.
Next, the data processing flow up to calculation of characteristic values will be described with reference to a flowchart.
In this processing example, a method is employed that successively stores data in a discharge period of time into array variables. Therefore, it is necessary to ensure a large memory area. This method can be easily realized by the software on a personal computer, but has difficulty in applying to a device with a small capacity memory, such as a PLC (Programmable Logic Controller, i.e. a sequencer). A second data processing example using no array variables is suitable for PLC.
When the processing of the flowchart of
Then, in step S103, raw data is input (x=data). Then, in step S104, the discharge start is judged. The discharge start is judged based on whether or not a discharge command is ON and x>xth. If it is judged that the discharge starts (Yes), step S105 starts counting a discharge time (count=count+1) and stores the data into an array (data(count)=x). Then, in step S106, the discharge end is judged. The discharge end is judged based on whether or not x<xth and count>0. While the end conditions are not satisfied, the loop of steps S103 to S106 is repeated. Each time the loop is repeated, a value of raw data x is stored into an array variable data( ). When the discharge ends (when x<xth and count>0), the processing goes out of the loop and enters a characteristic value calculation routine.
In step S107 of the characteristic value calculation routine, the length of the discharge time is judged (count>10) and, when the discharge time is long, the processing proceeds to step S108 where data corresponding to t1 from the beginning of the discharge data and data corresponding to t2 from the last of the discharge data are removed and then the characteristic values (average value avg, maximum value max, minimum value min, standard deviation dev, and discharge time t) are calculated.
avg=average{data(t1)˜data(t2)}
max=max{data(t1)˜data(t2)}
min=min{data(t1)˜data(t2)}
dev=stdev{data(t1)˜data(t2)}
t=count*sampling rate
On the other hand, when the discharge time is short, the processing proceeds to step S109 where the characteristic values are calculated based on the data up to the count value.
avg=average{data(t1)˜data(count)}
max=max{data(t1)˜data(count)}
min=min{data(t1)˜data(count)}
dev=stdev{data(t1)˜data(count)}
t=count*sampling rate
Then, in either case, in step S110, the calculated characteristic value data (avg, max, min, dev, t) are output and, in step S111, the variables are initialized. Then, the processing returns to step S103.
data(*)=0,count=0
avg=max=min=dev=t=0
This flowchart shows the processing example using no array variables. Steps S201, S202, and S203 are the same as steps S101, S102, and S103 in the flowchart of
When the data x exceeds xth, the judgment in step S204 becomes YES so that the processing enters a loop 2 to calculate respective characteristic values in sequence.
That is, step S205 counts up a discharge time (count=count+1) and substitutes the data (data=x) and steps S206 and S207 update the maximum value (when data>max, max=data is set). Further, next steps S208 and S209 update the average value/variance (when count>t1, totalx=totalx+data, totalx2=totalx+data^2, and avg=totalx/(count−t1) are set).
Further, steps S210 and S211 perform calculation of the standard deviation (when count>t1+1, dev=Sqr((totalx2−(count−t1)*(avg^2))/(count−t1−1)) is calculated). Further, steps S212 and S213 update the minimum value (when data<min, min=data is set). Then, step S214 calculates the discharge time (t=count*sampling rate).
When the discharge ends, the processing again enters the loop 1, outputs the results (step S217), initializes the variables (step S218), and repeats the loop 1 until the next discharge occurs. The discharge end is judged to be positive when data x<threshold value xth and count>0 are both satisfied and, when these conditions are satisfied (i.e. immediately after the discharge end), the data are once output to a database.
This flowchart shows the processing example corresponding to a case where the discharge is temporarily interrupted due to arc or the like, which is obtained by partly changing the second data processing example of
In this example, a discharge ON command from a PLC is input and conditions are branched. Discharge interruption is judged to be positive based on the condition that the discharge command is ON and data is less than a threshold value in steps S304A and S304B. In this case, the processing enters a loop 3 and does not count a discharge time or perform sequential calculation of data. When the discharge starts again, the judgment in step S304B becomes YES so that the processing enters a loop 2 to normally perform counting of the discharge time and sequential calculation of data.
Through the calculation processing performed in the calculation section 100 as described above, the proper characteristic values are calculated.
The data processing illustrated in each of
In the foregoing embodiments, the description has been given of the case where this invention is applied to the magnetic disk manufacturing systems. However, this invention is also applicable to other processing devices for processing substrates, for example, a semiconductor manufacturing system and so on. Further, this invention is applicable not only to a processing device for processing workpieces by sputtering, but also to a system that performs other processing such as CVD.
| Patent | Priority | Assignee | Title |
| 10316823, | Mar 15 2017 | Inventus Holdings, LLC | Wind turbine group control for volant animal swarms |
| 10620600, | Apr 07 2011 | ENDRESS+HAUSER SE+CO KG | Apparatus and system for determining, optimizing or monitoring at least one process variable |
| Patent | Priority | Assignee | Title |
| 6338777, | Oct 23 1998 | International Business Machines Corporation | Method and apparatus for sputtering thin films |
| 6741941, | Sep 04 2002 | Hitachi, Ltd. | Method and apparatus for analyzing defect information |
| 6772034, | Jul 12 2001 | GLOBALFOUNDRIES U S INC | System and software for data distribution in semiconductor manufacturing and method thereof |
| 6871112, | Jan 07 2000 | GLOBALFOUNDRIES Inc | Method for requesting trace data reports from FDC semiconductor fabrication processes |
| 6920405, | Sep 15 1999 | Becton, Dickinson and Company; Becton, Dickson and Company | Systems, methods and computer program products for constructing sampling plans for items that are manufactured |
| 7580768, | Sep 22 2006 | Texas Instruments Incorporated | Method of adjusting process variables in a processing flow |
| 7831326, | May 04 2007 | Applied Materials, Inc. | Graphical user interface for presenting multivariate fault contributions |
| 20010050220, | |||
| 20030183509, | |||
| 20040026235, | |||
| 20040182696, | |||
| 20040182697, | |||
| 20050159835, | |||
| 20050165731, | |||
| 20050177269, | |||
| 20050187649, | |||
| 20050205413, | |||
| 20050236266, | |||
| 20050278597, | |||
| 20060071673, | |||
| 20060167653, | |||
| 20060184264, | |||
| 20070038889, | |||
| 20070102291, | |||
| 20080021664, | |||
| 20080082294, | |||
| 20090292506, | |||
| CN1362660, | |||
| EP1220529, | |||
| JP11162851, | |||
| JP11316950, | |||
| JP2002294444, | |||
| JP2004186445, |
| Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
| Mar 30 2007 | Hoya Corporation | (assignment on the face of the patent) | / | |||
| Sep 25 2008 | KOBAYASHI, TAKUMI | Hoya Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 021612 | /0617 | |
| Sep 25 2008 | KOBAYASHI, TAKUMI | HOYA MAGNETICS SINGAPORE PTE LTD | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 021612 | /0617 | |
| Sep 15 2011 | HOYA MAGNETICS SINGAPORE PTE LTD | Hoya Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 026943 | /0593 |
| Date | Maintenance Fee Events |
| Apr 03 2018 | M1551: Payment of Maintenance Fee, 4th Year, Large Entity. |
| Mar 30 2022 | M1552: Payment of Maintenance Fee, 8th Year, Large Entity. |
| Date | Maintenance Schedule |
| Oct 14 2017 | 4 years fee payment window open |
| Apr 14 2018 | 6 months grace period start (w surcharge) |
| Oct 14 2018 | patent expiry (for year 4) |
| Oct 14 2020 | 2 years to revive unintentionally abandoned end. (for year 4) |
| Oct 14 2021 | 8 years fee payment window open |
| Apr 14 2022 | 6 months grace period start (w surcharge) |
| Oct 14 2022 | patent expiry (for year 8) |
| Oct 14 2024 | 2 years to revive unintentionally abandoned end. (for year 8) |
| Oct 14 2025 | 12 years fee payment window open |
| Apr 14 2026 | 6 months grace period start (w surcharge) |
| Oct 14 2026 | patent expiry (for year 12) |
| Oct 14 2028 | 2 years to revive unintentionally abandoned end. (for year 12) |